Snowflake Block

The Snowflake block connects to your Snowflake Data Warehouse to execute SQL queries, insert/update/delete data, manage tables, and run advanced data operations. Reach for it when a workflow needs to read from or write to Snowflake, with support for dynamic schema detection, parameterized queries, bulk loads, and connection pooling.

Overview

When to Use

• Query a Snowflake warehouse and feed results into downstream blocks.
• Insert, update, or delete rows in Snowflake tables from a workflow.
• Create, drop, or truncate tables, or run stored procedures.
• Bulk load data from a stage using CSV, JSON, Parquet, Avro, or XML formats.
• Inspect schema info or list tables for dynamic, schema-aware automations.
• Test connectivity to a Snowflake account before running heavier operations.

Configuration

Connection Mode

How to authenticate to Snowflake. Options:

Defaults to credentials. The relevant credential fields appear or hide based on this selection.

Account Name

Your Snowflake account identifier (e.g. mycompany.eu-central-1). Required.

Database

The Snowflake database to connect to. Required.

Warehouse

The virtual warehouse to use for query execution. Required.

Schema

The schema within the database. Required.

Username

Username for authentication. Required when connectionMode is credentials.

Password

Password for authentication (stored securely). Required when connectionMode is credentials.

Role

Optional Snowflake security role to assume for the session.

OAuth Token

OAuth token for authentication. Required when connectionMode is oauth.

Private Key

RSA private key for key pair authentication (stored securely). Required when connectionMode is keypair.

Private Key Passphrase

Passphrase for an encrypted private key. Only shown when connectionMode is keypair.

Keep Session Alive

Toggle to keep the Snowflake session alive indefinitely. Defaults to off.

Connection Timeout (seconds)

How long to wait when establishing a connection before failing. Range: 30–300 seconds. Default: 60.

Query Timeout (seconds)

Maximum time allowed for a query to run before it is aborted. Range: 30–3600 seconds. Default: 300.

Operation

The action to perform. Required. Options:

SQL Query

The SQL statement to execute. Shown for query, create_table, drop_table, and truncate operations. Required for those operations. Includes an AI wand that generates Snowflake-correct SQL (CTEs, window functions, VARIANT type, etc.) from a plain-English prompt.

Table Name

Target table (table_name or schema.table_name). Required for insert, update, delete, drop_table, truncate, and bulk_load.

Data Mode

How data is provided for insert and update operations. Options:

Insert Data

JSON object or array of objects representing the rows to insert. Shown for insert when dataMode is json.

Update Data (SET clause)

JSON object whose keys are column names and values are the new values. Shown for update when dataMode is json.

WHERE Clause

Filter condition for update and delete operations (e.g. id = 123 AND status = 'active').

Stage Name

Snowflake stage reference for bulk load (e.g. @my_stage or @%table_name). Shown for bulk_load.

File Format

File format of the data in the stage. Shown for bulk_load. Options:

Procedure Name

Name of the stored procedure to call. Shown for procedure.

Procedure Parameters

JSON array of positional arguments passed to the stored procedure (e.g. ["param1", 123]). Shown for procedure.

Use Parameterized Query

Toggle to use parameterized (bind-variable) queries for security. Shown for query, insert, update, and delete.

Query Parameters

JSON object of parameter bindings for the parameterized query (e.g. {"param1": "value1"}). Shown when useParameterizedQuery is on.

Output Format

How query results are returned. Shown for query, schema_info, and list_tables. Default: json. Options:

Enable Pagination

Toggle to paginate large result sets. Shown for query.

Page Size

Number of rows per page. Range: 10–10000. Default: 1000. Shown when pagination is on.

Page Number

Which page to retrieve (1-based). Range: 1–1000. Default: 1. Shown when pagination is on.

Enable Retry

Toggle to automatically retry failed operations.

Max Retries

Maximum retry attempts. Range: 1–10. Default: 3. Shown when retry is on.

Retry Delay (ms)

Milliseconds to wait between retry attempts. Range: 100–10000. Default: 1000. Shown when retry is on.

Enable Connection Pooling

Toggle to reuse database connections across invocations for better performance.

Max Connections

Maximum number of connections in the pool. Range: 1–50. Default: 10. Shown when pooling is on.

Enable Schema Detection

Toggle to automatically detect table schemas. Shown for insert, update, and schema_info.

Schema Table

Table name to run schema detection against. Shown when schema detection is on.

Enable Query Logging

Toggle to log all executed queries for debugging purposes.

Enable Metrics

Toggle to collect execution metrics and performance statistics.

Inputs & Outputs

• Inputs:

  • accountName (string) — Snowflake account name
  • database (string) — Database name
  • warehouse (string) — Virtual warehouse name
  • schema (string) — Schema name
  • username (string) — Username for authentication
  • password (string) — Password for authentication
  • role (string) — Security role to assume
  • connectionMode (string) — Authentication method
  • oauthToken (string) — OAuth token for authentication
  • privateKey (string) — Private key for key pair authentication
  • privateKeyPassphrase (string) — Private key passphrase
  • clientSessionKeepAlive (boolean) — Keep session alive
  • connectionTimeout (number) — Connection timeout in seconds
  • queryTimeout (number) — Query timeout in seconds
  • operation (string) — Operation to perform
  • sqlQuery (string) — SQL query to execute
  • tableName (string) — Target table name
  • dataMode (string) — Data input format
  • insertData (json) — Data to insert
  • updateData (json) — Data to update
  • whereClause (string) — WHERE clause for update/delete
  • stageName (string) — Stage name for bulk operations
  • fileFormat (string) — File format for bulk load
  • procedureName (string) — Stored procedure name
  • procedureParams (json) — Procedure parameters
  • useParameterizedQuery (boolean) — Use parameterized queries
  • parameters (json) — Query parameters
  • outputFormat (string) — Output format
  • enablePagination (boolean) — Enable result pagination
  • pageSize (number) — Number of rows per page
  • pageNumber (number) — Page number to retrieve
  • enableRetry (boolean) — Enable retry on failure
  • maxRetries (number) — Maximum number of retries
  • retryDelay (number) — Delay between retries in milliseconds
  • enableConnectionPooling (boolean) — Enable connection pooling
  • maxConnections (number) — Maximum number of pooled connections
  • enableSchemaDetection (boolean) — Enable automatic schema detection
  • schemaTable (string) — Table for schema detection
  • enableQueryLogging (boolean) — Enable query logging
  • enableMetrics (boolean) — Enable performance metrics

• Outputs:

  • success (boolean) — Whether the operation was successful
  • data (json) — Query results or operation data
  • rowCount (number) — Number of affected or returned rows
  • executionTime (number) — Query execution time in milliseconds
  • queryId (string) — Snowflake query ID for tracking
  • error (string) — Error message if the operation failed
  • schema (json) — Table schema information
  • tables (json) — List of tables in the database
  • columns (json) — Column information for the table
  • hasNextPage (boolean) — Whether there are more pages available
  • totalPages (number) — Total number of pages
  • currentPage (number) — Current page number
  • connectionInfo (json) — Connection status and details
  • warehouseInfo (json) — Warehouse utilization information
  • metrics (json) — Performance metrics and statistics
  • queryLog (json) — Query execution log
  • loadStatus (json) — Bulk load operation status

Tools

Snowflake Data Warehouse (snowflake_connector) — Connects to Snowflake via the snowflake-sdk, establishes a connection (credentials, OAuth, or key pair), executes the requested SQL operation, and returns structured results including row data, row count, execution time, query ID, schema info, pagination metadata, and performance metrics. All parameters from the block are forwarded directly to this tool via a POST request to /api/tools/snowflake.

YAML Example

snowflake_1:
  type: snowflake
  name: "Snowflake"
  inputs:
    connectionMode: "credentials"
    accountName: "mycompany.eu-central-1"
    database: "ANALYTICS"
    warehouse: "COMPUTE_WH"
    schema: "PUBLIC"
    username: "{{SNOWFLAKE_USER}}"
    password: "{{SNOWFLAKE_PASSWORD}}"
    operation: "query"
    sqlQuery: "SELECT * FROM customers LIMIT 100;"
    outputFormat: "json"
  connections:
    outgoing:
      - target: next-block-id